Mathematics Colloquia and Seminars

Return to Colloquia & Seminar listing

Mathematical Methods Applied to the Representation and Visualization of Large Scientific Data Sets

Student-Run Research Seminar

Speaker: Bernd Hamann, UC Davis
Location: 693 Kerr
Start time: Wed, Oct 19 2005, 12:10PM

Established data compression and multiresolution techniques are an essential part of the processing and exploration of massive scientific data sets. Most existing techniques, especially wavelet-based approaches, assume that data sets to be represented hierarchically are defined on rather regular underlying mesh structures. This talk is aimed at pointing out some directions for establishing wavelet-based multiresolution methods for scientific data sets whose underlying mesh structure is much less restrictive. The possible directions that are illumed to are based on concepts from approximation and spline theory, computational geometry, and optimization. Further, a close coupling between driving application domains and supporting data approximation and visualization technology is emphasized.